Early problem drinkers (i.e., those whose heavy drinking has either begun to cause them alcohol-related harm or puts them at risk of such harm) are highly prevalent in the United States. These patients often visit primary care physicians who, increasingly, are learning to identify and intervene with this patient group. Recent, controlled trials have shown that brief interventions in the primary care setting can reduce drinking and alcohol-related problems in patients who are without substantial evidence of alcohol dependence. Although naltrexone has been approved for the treatment of alcohol dependence, few pharmacotherapy studies have been undertaken with early problem drinkers. We propose an eight-week trial of naltrexone vs. placebo, combined with coping skills treatment that either focuses on targeted use of medication or serves as background to daily use of the medication. One hundred sixty early problem drinkers recruited through screening in primary care medical settings will be randomly assigned to one of four treatment groups (targeted naltrexone/daily naltrexone/targeted placebo/daily placebo). Life events, mood, desire to drink, perceived self-efficacy, and drinking will be monitored through daily self-report questionnaires. Follow-up evaluations will be conducted at the end of treatment and again three, six, and twelve months later. We hypothesize that 1) during treatment naltrexone will result in greater reductions in the desire to drink and in alcohol consumption and related problems; 2) after treatment, an advantage will emerge for the targeted medication-coping skills group; and 3) the group receiving targeted naltrexone will have the best outcomes during the follow-up period, due to the interaction of active medication and skills training, which are expected to act synergistically to enhance perceived self-efficacy. Specific mechanisms by which naltrexone exerts its beneficial effects will be examined by means of the daily process measures.